EpiCompare compares different epigenetic datasets for quality control and benchmarking purposes. The report is divided into three sections:

  1. General Metrics: Metrics on fragments (duplication rate) and peaks (blacklisted peaks and peak widths) of input files
  2. Peak Overlap: Percentage and statistical significance of overlapping and unon-overlapping peaks.
  3. Functional Annotation: Functional annotation (ChromHMM, ChIPseeker and enrichment analysis) of peaks.


Input Datasets

Total of 6 peak files:

## [1] "File1: active_seacr"
## [1] "File2: active_macs"
## [1] "File3: H3K27me3_seacr"
## [1] "File4: H3K27me3_macs"
## [1] "File5: CnT_seacr"
## [1] "File6: CnT_macs"

1. General Metrics

Peak Information

Column Description:

  • PeakN before tidy: Total number of peaks including those blacklisted and those in non-standard chromosomes.

  • Blacklisted peaks removed (%): Percentage of blacklisted peaks present in the sample. ENCODE blacklist includes regions in the hg19 genome that have anomalous and/or unstructured signals independent of the cell-line or experiment.

  • Non-standard peaks removed (%): Percentage of peaks identified in non-standard and/or mitochondrial chromosomes. Identified using BRGenomics::tidyChromosomes().

  • PeakN after tidy: Total number of peaks after filtering blacklisted peaks and those in non-standard chromosomes.

    NB: All analyses in EpiCompare are conducted on tidied datasets (i.e. blacklisted peaks and those in non-standard chromosomes removed)


Sample PeakN before tidy Blacklisted peaks removed (%) Non-standard peaks removed (%) PeakN after tidy
active_seacr 3211 17.80 4.390 2497
active_macs 2526 23.20 7.050 1762
H3K27me3_seacr 107569 2.93 0.435 103953
H3K27me3_macs 89900 2.85 0.414 86965
CnT_seacr 6669 3.42 0.390 6415
CnT_macs 13456 3.47 0.498 12922

Fragment Information

Metrics on fragments is shown only if Picard summary is provided. See manual for help.

Column Description:

  • Mapped_Fragments: Number of mapped read pairs in the file.
  • Duplication_Rate: Percentage of mapped sequence that is marked as duplicate.
  • Unique_Fragments: Number of mapped sequence that is not marked as duplicate.



Peak widths

Distribution of peak widths in each sample.



2. Peak Overlap

Percentage Overlap

Heatmap of percentage of overlapping peaks between samples. Hover over the heatmap for percentage values.



Statistical Significance

The plot is shown only if a reference peak file is provided and stat_plot = TRUE. Depending on the format of the reference file, EpiCompare outputs different plots:

  • Reference dataset has BED6+4 format (peakcalling performed with MACS2): EpiCompare generates paired boxplot per sample showing the distribution of -log10(q-value) of reference peaks that are overlapping and non-overlapping with the sample dataset.
  • Reference dataset does not have BED6+4 format: EpiCompare generates a barplot of percentage of overlapping sample peaks with the reference, coloured by statistical significance (adjusted p-value) of the overlap.



3. Functional Annotation

3.1 ChromHMM

ChromHMM annotates and characterises peaks into different chromatin states. The annotation used in EpiCompare were obtained from here. f

All samples

ChromHMM annotation of individual samples.

Overlap: Sample peaks in Reference peaks

ChromHMM annotation of sample peaks found in reference peaks.

Overlap: Reference peaks in Sample peaks

ChromHMM annotation of reference peaks found in sample peaks.

Unique: Sample peaks not in Reference peaks

ChromHMM annotation of sample peaks not found in reference peaks.

Unique: Reference peaks not in Sample peaks

ChromHMM annotation of reference peaks not found in sample peaks.

3.2 ChIPseeker

EpiCompare uses annotatePeak function in ChIPseeker package to annotate peaks with the nearest gene and genomic region where the peak is located. The peaks are annotated with genes taken from the annotations of human genome hg19 provided by Bioconductor.

3.3 Functional Enrichment Analysis

EpiCompare performs KEGG pathway and GO enrichment analysis using clusterProfiler. annotatePeak function in ChIPseeker package is first used to assign peaks to nearest genes and biological themes amongst the genes are identified using ontologies (KEGG and GO). The peaks are annotated with genes taken from the annotations of human genome hg19 provided by Bioconductor.

KEGG

GO